Abduction

Logic is the systematic study of cogent reasoning. Reasoning emphasizes the process of drawing inferences (conclusions) from some initial information (premise). In standard logic, when the truth of the conclusions follows the premises, we refer to it as deduction. That is, the truth of the premises guarantees the truth of the conclusions. However, when the truth of the conclusion is rendered to be more credible but not necessarily guaranteed, we refer to it as induction. Then there is abduction, a process yielding an explanation of an event or set of events. Reasoning, along with thinking, has been at the core of modern science. Let’s elaborate on these concepts.

Deduction is generally the process of concluding by reasoning. Deductive reasoning has been the most common method of reasoning in the natural sciences for more than 150 years. The deductive method is also called the hypothetico-deductive method. Aristotle defined it as going from the general to the specific. The researcher first formulates a theory or set of theories which leads to the formation of hypotheses that are then tested against the data for their predictive success. It is the process of moving from a general theory to a specific set of conclusions.

An example of deductive reasoning is if A=B (premise), and B=C (premise), then A=C (conclusion). If I need to be at my office at 9:30 am and it takes 30 minutes to drive, I can use deductive reasoning that conclude that I need to leave by 9 am at the latest. Deduction is generally non-ampliative and certain. Non-ampliative means the conclusion does not go beyond what is given in the premises. Certain means, when the premises are true the conclusion must be true. However, much of the world that we observe cannot be summed up in a neat deductive argument. It is difficult to find completely true premises.

Induction or the inductive method of scientific reasoning is when systematic observations lead to the development of theories. Inductive reasoning is the process of moving from specific observations to generalizations. In this type of research, researchers first make observations or collect facts, look for patterns, commonalities, or themes in the observations (upward hypothesis), and then develop a theory that can explain the phenomenon observed. The first part is often referred to as enumerative induction while the later part of generalization is referred to as induction by intuition. However, in deduction, where the truth of the conclusion is guaranteed in indiction it is only probable. For example, when you try to start the car but it does not start the engine, you would be inclined to think that your battery might be dead. While it is certainly possible that is no guarantee that this conclusion is true. There could be other reasons that might be stopping the car from starting.

Since the process of induction draws conclusions that go beyond the content of its premises it is ampliative which makes the conclusions probable rather than certain. That is, even when the truth of the premises is taken for granted, the conclusion is only likely. Therefore it must be subjected to further testing. For example, I might know the shortest way to reach from my home to the office. But I also know from experience that the shortest way is not always the fastest way to reach my office depending on the traffic and time of the day. So even when the premises are true the conclusion is not always true.

However, depending on the time of the day I can make an educated guess of what would be the fastest way to reach my office. This is referred to as abduction. Abduction does not reason from premises to a conclusion. Rather it deals with ruling out possible conclusions till we are left with the most plausible one given the evidence. It involves drawing a conclusion based on what is known.

Abduction, defined by C. S. Peirce, is the process of forming an explanatory hypothesis from an observation requiring an explanation (Peirce, 1958, volume 5 para 171). It is sometimes also referred to as inference to the best explanation. When we encounter new phenomena that cannot be explained by the premises at hand, we pick out certain characteristic features of this new phenomenon and find relationships among these features. We can then form various hypotheses that explain this new phenomenon and use our experience to select the most plausible explanation and eliminate implausible explanations.

For example, route A is the shortest route to my office from my home (premise) and morning 8 am to 10 am is the when there is high traffic on route A (premise). I left my home at 9 am and I was at the 9:30 am meeting at my office. I must have taken an alternative route to my office (conclusion). Here using only the premises we cannot prove this conclusion based on deductive or inductive reasoning. But given what we know from the premises and combining it with our experience, taking an alternative route is the most plausible explanation of the events, given that it takes about 30 minutes to reach my office when I use route A. Here the conclusion is only a likely or probable explanation for the events or the best inference to an explanation. It only uses the information at hand.

Abductive reasoning is a creative way of forming hypotheses to test. Abduction is an inferential process that yields causal explanations as we reason backward from effect to cause that explains it. Peirce identified abduction as the first step of scientific inquiry which, according to him, follows abduction, deduction, and induction. We use abductive reasoning all the time making educated guesses. Sherlock Holmes was considered a master of using abductive reasoning. Abduction relies on knowledge of causal relationships and uses temporally ordering of the set of possibilities. Abductive reasoning is also an exercise in imagination, in that, abduction can create ideas with propositional content which can then be verified. Abduction, thus, has the potential to create new knowledge and explanations. So much of Einstein’s work was carried out as a thought process that went beyond induction and deduction. Abductive reasoning is also at the core of the rise of artificial intelligence. AI uses abduction and deduction and cycles around from one to another till it solves the problem.

Abductive reasoning plays a central role in many fields from from AI and visual cognition to neuroscience, medicine and mathematics to architecture and design, economics, education, and human sciences.

Abductive reasoning is strengthened when the generated hypothesis, which can provide the best inference to an explanation, has qualities like simplicity, breadth, consistency with prior knowledge, and coherence. That is, how the quality of an explanation influences the degree of belief in that explanation as well as belief in what is explained.

The creation of scientific theories is an interplay between induction, abduction, and deduction that we also observe in our everyday lives. Abductive reasoning is one of the three fundamental types of reasoning used in science along with inductive and deductive reasoning.

Sources

Johnson-Laird, P. (2008). How we reason. Oxford University Press.

Magnani, L. (2001), Abduction, reason, and science: Processes of discovery and explanation. New York: Kluwer Academic/Plenum Publishers.

Peirce, C. S. (1958). Collected papers of Charles Sanders Peirce. Harvard University

Peirce, C. S. (1955). Abduction and induction. In J. Buchler (Ed.), Philosophical writings of Peirce (pp. 150–156). New York: Dover. [Original work published in 1903].

Thagard, P., Shelley, C. (1997). Abductive reasoning: Logic, visual thinking, and coherence. In: Dalla Chiara, M.L., Doets, K., Mundici, D., van Benthem, J. (eds) Logic and Scientific Methods. Synthese Library, vol 259. Springer, Dordrecht. https://doi.org/10.1007/978-94-017-0487-8_22